

***Replication Do-File for
***Do Voters Respond to Relative Economic Performance? Evidence from Survey Experiments
***by Selim Erdem Aytaç
***Forthcoming in Public Opinion Quarterly
***This version: February 2020


**Table 3
**Descriptive statistics of variables used in analyses

*Use UK Data

tabstat competence female age college social_grade region income p_cons, stat(n mean sd min max)

*Use Turkey Data

tabstat competence female age college employed income p_akp, stat(n mean sd min max)


**Table 4
**Average treatment effects in survey experiments

*Use UK Data

reg competence domestic international, robust
reg competence domestic international female age college social_grade region p_cons income, robust

*Use Turkey Data

reg competence domestic international, robust
reg competence domestic international female age college employed p_akp income, robust


**Figure 1
**Distribution of responses to outcome questions in the control condition

*Use UK Data

hist competence if domestic==0 & international==0, d percent

*Use Turkey Data

hist competence if domestic==0 & international==0, d percent


**Figure 2
**Average treatment effects – broken down by respondents’ views on the incumbent’s responsibility for economic outcomes

*Use UK Data

reg competence domestic international if pmcontrol==0, robust

estimates store b_1

reg competence domestic international if pmcontrol==1, robust

estimates store b_2

*Use Turkey Data

reg competence domestic international if pmcontrol==0, robust

estimates store t_1

reg competence domestic international if pmcontrol==1, robust

estimates store t_2


*Combine the estimates to reproduce the figure

coefplot (b_1, label(PM has no/little control over the economy)) (b_2, label(PM has some/significant control over the economy)), bylabel(British Survey)   ///
       || (t_1)           (t_2)         , bylabel(Turkish Survey)  ///
     ||, levels(90) drop(_cons) xline(0) byopts(xrescale)



**Figure 3
**Average treatment effects – broken down by respondents’ education levels among those who think the PM has at least some control over the economy

*Use UK Data

reg competence domestic international if pmcontrol==1 & college==0 , robust

estimates store b_3

reg competence domestic international if pmcontrol==1 & college==1 , robust

estimates store b_4

*Use Turkey Data

reg competence domestic international if pmcontrol==1 & college==0 , robust

estimates store t_3

reg competence domestic international if pmcontrol==1 & college==1 , robust

estimates store t_4

*Combine the estimates to reproduce the figure

coefplot (b_3, label(Respondents without college education)) (b_4, label(Respondents with college education)), bylabel(British Survey)   ///
       || (t_3)           (t_4)         , bylabel(Turkish Survey)  ///
     ||, levels(90) drop(_cons) xline(0) byopts(xrescale)
	 

*end of file
